Global macro lessons
They may seem worlds apart, but global macro managers might have some useful things to teach pension funds, writes Martin Steward
The highly regulated world of a European pension plan looks like a different planet from the freewheeling contrarianism of the global macro hedge fund manager. The former sits on a strategic portfolio of assets for three to five years, tinkering at the margins, based on time-honoured volatility and correlation metrics. The latter sits on cash and Treasuries, scouring the globe for and pouncing on a few asymmetric risk trades that can last for years or seconds.
And yet, Patrick Welton, CEO of Welton Investment Corporation, whose Global Directional Portfolio is a macro strategy with a more long-term bias than the classic style, suggests: "One could argue that a well-diversified macro manager looks something like what a pension fund is trying to be: multi-asset class, multi-strategic."
Shared ground, different methodologies: often a good place for lessons to be exchanged. These are four suggestions.
1. Liquidity, liquidity, liquidity
For a pension fund, portfolio construction and diversification is almost everything, whereas a global macro firm's traders run individual books that can become extremely concentrated, both in themselves and in aggregate. But that means risk management at position and trading-book level is rigorous. Typically, firms practice drawdown risk budgeting (successful traders get more risk budget, losers less) alongside disciplined stop-loss and profit-taking processes. Thomas Weber, head of hedge fund investment management at LGT Capital Partners, observes that some systematic macro traders post positive returns even as 60% of their positions lose money.
"Risk control involves monitoring traders' positions, cutting those that are going wrong, taking profits at the appropriate time when it's going right, and maintaining appropriate levels of risk exposure across the firm," says Ian Plendereith, chairman of BH Macro, the listed vehicle investing in Brevan Howard's flagship Master fund.
‘Appropriate' portfolio-level risk does not necessarily equate to ‘diversified'. "We have a handful of strategies," says Peter Allwright, a currency and rates specialist working on Threadneedle's Crescendo fund, "and generally they will all be correlated to some extent. The decision we make is, are we risk-on or risk-off, and how much risk do we want to take? Then we make sure we will have liquidity when we need it. In some ways ours is the easiest fund for our chief risk officer to look after because we are so liquid. Instead of fighting to survive, we survive to fight."
Risk management can focus on individual positions only because they are in the most liquid markets. As a result, tight portfolio-level risk limits can be applied with surprising rigour. "Among portfolio-level risk measures, perhaps the most intuitive is ‘How many dollars am I willing to lose?' " says Leon Diamond, portfolio manager on Castlestone Management's Porcupine fund. "I stress test assuming that everything in the portfolio correlates at 1.0, then I implement that worst-case scenario as a hard stop-loss on the portfolio. On average that stop sits at about 3-4%."
The lesson from all this? Portfolio-level risk management should not be insurance against lax position-level risk management - particularly if the main risk that is being taken in those positions, and possibly ignored, is liquidity risk.
2. Redefine your sources of risk
Still, pension funds are not short-term traders. We have to turn to the more asset allocation-minded macro managers for lessons in portfolio construction. "When it comes to optimisation, keep in mind the maxim ‘garbage-in, garbage-out'," warns Nicolai Borcher Hansen, founding partner and CIO at multi-asset manager AROS Capital Partners. "When entering correlations, do you really believe they can be true?"
In this regard, Welton speaks the long-term allocators' language. A strategic allocation should be "for all time", he says, but most pension funds "simply aren't diversified enough" to achieve this because they use legal structures (stocks and bonds) to construct investment hypotheses. Instead, macro managers use "functional asset classes". He says: "One can classify an investment's valuation response to increasing or decreasing demands for factors such as liquidity, volatility, quality, inflation or cashflow production, among others, and use those to build portfolios that do well over all times: when equity beta and multiples rise or fall; when liquidity demands rise or fall; when there's a flight to or from quality; when there is inflation or deflation; or by picking up as many synthetic cashflows as it can."
Dori Levanoni, co-head of global macro at First Quadrant, describes a similar approach. Asset classes are not betas but collections of betas: even sovereign debt wraps up ‘safe-haven' beta with its basic, cost-of-capital beta. "Some of these betas have fat tails, some thin tails, some negative skew, and so on, and the weighting of each piece within an asset changes through time," he says. These non-normal distributions contaminate risk management, so the macro manager aims to separate them from alpha and then dynamically trade them to look more like alpha - mesokurtic, with a slight positive skew.
For a pension fund, of course, these betas are core, and exposure to them is constructed to match liabilities. The lessons? Recognise, when constructing strategic allocations, that the historical tracking error of equities or bonds against your liabilities reveals little about the potential correlation of "functional asset classes", or betas, with those liabilities; and that tactically trading those betas can dampen asset-liability mismatches resulting from their non-normal distributions.
3. Try to engineer optionality
Alternatives to tactical trading are strategic optionality and convexity - Welton describes portfolios that perform irrespective of market direction. Global macro funds engineer convexity by being liquid enough to follow markets. That is impracticable for most of a pension fund's portfolio, but optionality can also be synthesised through position construction. Diamond describes participating in the long AUD/JPY carry trade: he knew it represented extreme left-tail risk in the event of a sudden unwind, so he bought deep-out-of-the-money options that could be exercised in the event that such an unwind sucked liquidity from the market.
In addition he outlines a hedge fund stalwart in the shape of the beta extension: by going long the S&P500 but short the NDX100 and Russell2000 he is directionally short equity market beta (and especially tech and the US commercial banks that hold so many CMBS) while retaining exposure to large energy companies and banks that he expects to benefit from strong oil prices and the steep yield curve. Welton offers the example of a beta-neutral basket of equity index futures, which he argues should rise in value as correlations rise, but exhibit mixed, close-to-zero returns through "normal" times: "That gives you optionality on a rise in correlations, which would be negatively correlated to equity beta," he says.
4. We can't all be macro funds
Unfortunately, all of this involves alpha or radical reshaping of beta - and therefore the limiting factor of macro-consistency. The uniqueness of each pension scheme's liability stream offers some scope for this to work at the aggregate level: an ageing miners' pension scheme could trade with or act as counterparty to a younger tech company's scheme, for example. Pareto efficiency (where both parties benefit from an exchange) can exist before market efficiency.
But the constraint is suggested by the lack of Pareto efficiency between global macro funds and pension schemes. "Global macro feeds off the inefficiency of institutional asset allocators," notes Kevin Arenson, CIO at fund of funds Stenham Advisors. Global macro is usually about taking asymmetric risks, which are necessarily contrarian. At most points in time a pension fund will necessarily have a higher beta to the market portfolio than a high-alpha global macro fund, and the bigger the pension fund the higher that beta will average out.
"A big pool of money is necessarily the crowd," says Alex Allen, CIO at fund of funds Eddington Capital Management. "Can pension funds learn anything from global macro funds? Sure. And individual pension funds are free to do what they want, of course. But if they all start pursuing one or more of the trades or methods of global macro funds, global macro will move on to something else. It would be like a dog chasing its tail."
The adage has it that if you can't beat them, you should join them. Some of these lessons are applicable by institutional asset allocators, but ultimately the markets will always make them feed alpha to global macro funds. Paying 2-and-20 to claw some back might not seem like Pareto efficiency - and it isn't. But the deal would be much less efficient for the pension fund if it were merely paying for a steady, uncorrelated alpha. However, the convexity of a diversified global macro exposure makes it more likely to match a right-tail risk with a pension fund's left-tail risk. The deal resembles insurance against the nastiest economic outcomes - something to which Pareto might have been more favourably disposed.